blob: 10ae0129b88d97304b2f127f706121c997e92b25 (
plain) (
tree)
|
|
[
{
message: <<EOF
promtool: Changed rule numbering from 0-based to 1-based when reporting rule errors.
Added `prometheus_remote_storage_read_queries_total` counter and
`prometheus_remote_storage_read_request_duration_seconds` histogram, removed
`prometheus_remote_storage_remote_read_queries_total` counter.
Added buckets for longer durations to
`prometheus_remote_storage_sent_batch_duration_seconds` histogram.
WAL compression is enabled by default.
EOF
maximum_version: "2.20"
type: upgrade
}
{
message: <<EOF
The `prometheus_sd_kubernetes_cache_*` metrics were removed.
The `prometheus_sd_kubernetes_workqueue_latency_seconds` and
`prometheus_sd_kubernetes_workqueue_work_duration_seconds` now show correct
values in seconds.
The `query` label on `prometheus_remote_storage_*` metrics has been changed to
`remote_name` and `url`.
EOF
maximum_version: "2.15"
type: upgrade
}
{
message: <<EOF
The `prometheus_tsdb_wal_reader_corruption_errors` metric was renamed to
`prometheus_tsdb_wal_reader_corruption_errors_total`.
EOF
maximum_version: "2.11"
type: upgrade
}
{
message: <<EOF
The `prometheus_tsdb_storage_blocks_bytes_total` metric is now
`prometheus_tsdb_storage_blocks_bytes`.
EOF
maximum_version: "2.8"
type: upgrade
}
{
message: <<EOF
The `storage.tsdb.retention` flag has been deprecated in Prometheus 2.7.0.
Users of this flag will want to update their configuration to use the newer
`storage.tsdb.retention.time` flag.
EOF
maximum_version: "2.7"
type: upgrade
}
{
message: <<EOF
Users of `marathon_sd` should now use `auth_token` and `auth_token_file` for
token-based authentication instead of `bearer_token` and `bearer_token_file`
respectively.
Metric names for the HTTP server metrics were changed.
EOF
maximum_version: "2.3"
type: upgrade
}
{
message: <<EOF
The `file_sd` mtime metric was renamed.
EOF
maximum_version: "2.2"
type: upgrade
}
]
|
